I was looking for replacement lightswitches; the old ones are probably 25 years old, some are broken. I got a new one that looked more or less the same, but I got home and realized it doesn't have the same feel as the old switch. And it isn't symmetric - it has a different click "on" vs "off".

Here's the thing - if you take your comment to the logical limit, the only engineering requirement of a product is to make the sale. With this switch, I took it home and installed it, and probably will not return it, so even if I decide to throw it in the trash, it wasn't "underengineered".

But everything is becoming like this, and so people and companies that "irrationally" overengineer stuff are becoming more valuable, as are their brands. Even if you don't care about what I care about, there are too many factors in any product for either of us to track. So there is always incredible value in someone that pays attention to the aspects we are not paying attention to. But that is overengineering. There is no alarm bell that tells you you have crossed over from optimizing what your customer cares about to screwing them over with a piece of crap. Assuming anyone cared.

If you optimize everything, you destroy your brand.

Sure, they didn’t engineer for the feel of the click, and you are evidence that there’s a market for it. Car manufacturers get this — flip the windshield wiper lever on a S-Class Mercedes and it feels completely different from doing it on a Kia. But that wouldn’t be overengineering if there is a real customer need.

If you want a low cost and low quality product you are free to purchase it. But many people prefer long-lasting, reliable products and there's a market for that.

You don't want to put your entire family in a car with poor quality control that will let me down in the middle of nowhere during winter.